1. Field of the Invention
The invention relates to and has among its objects the provision of hybridomas that produce and secrete monoclonal antibodies which have high affinity for Bowman-Birk inhibitor, immunoassay methods for the determination of Bowman-Birk inhibitor utilizing the monoclonal antibodies, and immunoaffinity techniques for attaching Bowman-Birk inhibitor to a solid phase.
2. Description of the Art
The protein of legumes, especially soybeans (Glycine max), is widely used in human foods. Soy protein is used in a variety of forms including infant formulas, tofu, soy protein isolates, soy flour, textured soy fibers, and soy sauce. Soybean protein products, properly processed, serve as an excellent source of low cost, high quality protein for human needs. Soybeans are even more widely used as a component of animal feeds and are a major export commodity.
Protease inhibitors occur widely in the legume family, comprising 5 to 10% of the storage proteins of the seed. Bowman-Birk inhibitor (BBI) encompasses a class of low molecular weight, double-headed inhibitors (that is, they have two reactive sites which bind to and inhibit proteases). BBI's are products of a group of similar genes and some are also proteolytically modified. In the soybean, the two principal protease inhibitors are the Kunitz trypsin inhibitor (KTI) and the classical Bowman-Birk inhibitor, which inhibits both trypsin and chymotrypsin. Recent evidence suggests that dietary protease inhibitors such as BBI may have a beneficial anticarcinogenic effect through their interaction with a cellular serine protease (J. Yavelow et al., Cancer Research (Suppl.) 43: 2454s-2459s (1983); J. Yavelow et al., Proceedings of the National Academy of Sciences USA 82: 5395-5399 (1985)). On the other hand, animal studies suggest that active (undenatured) protease inhibitors may be toxic to humans, and may adversely affect nutritional quality. Because protease inhibitors have both beneficial and adverse effects, a need exists to determine the amount and nature of active protease inhibitors present in foods and animal feeds, and determine the balance in food which will result in optimum benefit. In addition, methods for measuring protease inhibitors in plant tissues are needed for evaluating the results of breeding and recombinant DNA studies.
One method to minimize the adverse effects of protease inhibitors in foods is to reduce the amount of active inhibitor. Approaches to reduction of active protease inhibitor content are physical and chemical treatment of soy products and genetic alteration of the soybean crop. Although the protease inhibitor activity is largely inactivated by denaturation through conventionally applied heat treatments of soy flour, 10 to 15% residual activity remains (J. J. Rackis et al., in M. Friedman (ed.) Nutritional and Toxicological Significance of Enzyme Inhibitors in Foods, 299-347, Plenum Press, New York (1986)). The problems with this approach are that the nature of the residual activity is difficult to characterize, and the process is costly in energy usage. Genetic modification of the soybean plant to develop varieties with low protease inhibitor content has inherent limitations. Desirable nutritional value and potential anticarcinogenic activity may be lost concomitant with the reduction of protease inhibitor content. Cross-pollination of the genetic variant by a wild-type cultivar could result in reexpression of protease inhibitor genes. In addition, some protease inhibitors may function to prevent predation of the crops by insects, so that some protease inhibitory activities may be important for the agronomic value of soybean cultivars. With either genetic or physicochemical approaches to the reduction of protease inhibitor activity, the ability to measure low levels of protease inhibitors in soybean tissues and derived food products is essential. Since the effects of plant-derived protease inhibitors are of increasing importance in human and animal nutrition, food safety, and plant genetics, it is important to establish the exact protease inhibitor composition of a sample. What is needed is an assay for BBI which has the following characteristics: (1) it is sufficiently sensitive and accurate to measure the low levels of BBI that are present in processed foods, (2) it can differentiate between active BBI and denatured BBI in processed samples, (3) it can differentiate BBI from KTI, and optionally (4) it can differentiate an individual BBI from other BBI's.
Current assay techniques are not capable of providing this information. With regard to enzyme assays, because both BBI and KTI exist as several isoforms, which are derived from different genes or are produced by proteolysis, it is impossible to establish the exact protease inhibitor composition of a sample through enzymatic assay. Moreover, enzyme assays often give inaccurate results with processed samples having low residual activity as found in commercial foods due to inhibition of enzymatic activity or interference by nonprotein components. Further, enzymatic assays do not differentiate among the various specific and nonspecific inhibitors present.
Although polyclonal antibodies which bind BBI have been reported, these molecules appear to have low antigenicity. To overcome this problem, BBI crosslinked with glutaraldehyde was used to elicit an immune response. However, sensitive immunoassay methods, useful to measure low levels of BBI that are present in seeds, tissues, or complex samples including processed foods or feed could not be carried out using the prior art antibodies. With regard to antibodies to BBI that have been reported, a first group of researchers, E. Offir et al., Israel Journal of Chemistry 9: 17BC-18BC (1971) and Y. Birk, Bayer Symposium V "Proteinase Inhibitors, " Spring-Verlag, Berlin, pp. 355-361 (1974), described antibodies to classical BBI which crossreacted with lima bean inhibitor (LBI), (an inhibitor which is about 90% homologous to classical BBI) but had only slight binding to KTI. A second group, D. L. -R. Hwang et al., Biochimica et Biophysica Acta 495: 369-382 (1977) and D. L. Hwang et al., Plant Physiology 61: 30-34 (1978), described antibodies elicited with glutaraldehyde-treated BBI, a treatment which appeared necessary to elicit antibodies suitable for radioimmunoassay (RIA) of BBI. This group set up RIA's for BBI, using BBI iodinated with .sup.125 I as labeled ligand. The RIA relied on precipitation of the BBI-antibody complexes, did not use solid-phase methods, and did not address the selectivity for active BBI as opposed to denatured BBI. Most strikingly, the RIA was insensitive--0.1 mg BBI/ml was the approximate limit of sensitivity. The usefulness of such an assay would be limited to very concentrated solutions rich in BBI. The assay would not be useful for complex food samples, for example. The antibodies were used to distinguish classical BBI from other protease inhibitors in soybean extracts and to monitor the release of inhibitors upon germination of soybeans. A third group of researchers, A. L. Tan-Wilson and K. A. Wilson, Phytochemistry 21: 1547-1551 (1982); A. L. Tan-Wilson et al., Plant Physiology 70: 493-497 (1982), and A. L. Tan-Wilson, et al., Journal of Agricultural and Food Chemistry 33: 389-393 (1985) also used antibodies elicited with glutaraldehyde-treated BBI. They noted changes in antiserum specificity during the course of immunization of the rabbits, and used a precipitation method--radial immundiffusion--to measure changes in BBI during germination of soybeans and the concentration of BBI in different tissues of soybean. Only unlabeled BBI was used in these assays, and specificity for active BBI as opposed to denatured BBI or LBI, or applicability of methods to food samples was not addressed.
The production of monoclonal antibodies by fusion of spleen cells and myeloma cells has been described previously by G. Kohler and C. Milstein, Nature 256: 495-497 (1975) and many other investigators. Monoclonal antibodies against protease inhibitors from animal sources are known (See P. Herion et al., Bioscience Reports 4: 139-148 (1984)). Monoclonal antibodies to the plant protease inhibitor KTI, a single-headed inhibitor of trypsin (molecular weight about 20,000), have been described and applied to the measurement of KTI in soy samples (D. L. Brandon and M. Friedman, Abstract, 6th International Congress of Immunology, Toronto, Canada, July 6-11, 1986; D. L. Brandon et al., in M. Friedman (ed.) Nutritional and Toxicological Significance of Enzyme Inhibitors in Foods, 449-467, Plenum Press, New York (1986), and D. L. Brandon et al., Journal of Agricultural and Food Chemistry 35: 195-200 (1987)).
As stated above, previous investigators treated BBI with the crosslinking agent glutaraldehyde in order to render it sufficiently immunogenic. It was therefore unknown whether the appropriate monoclonal antibody specificity could be elicited with unmodified, low molecular weight (about 8000) BBI. It was also unknown whether antibodies could be generated with high affinity to BBI to provide for detection of low levels of BBI present in seeds, tissues, processed foods and the like. Further, it was unknown whether active BBI could be distinguished from BBI denatured by heat or chemical treatment or whether sufficiently specific antibodies could be prepared which could differentiate among classical BBI, and other BBI's such as LBI. It was also unknown whether immunoassays would yield linear responses useful in detecting low levels of BBI in samples containing multiple protease inhibitors and containing phytate, fat, fiber, and other potentially interfering substances. Further, it was unknown whether BBI could be measured in the presence of BBI-binding proteins such as chymotrypsin or other proteins which might be found in products and tissues derived from the soybean. The identification of BBI as part of a complex is important because it would allow quantitation under conditions where enzymatic assays would be inaccurate or impossible. Quantitation of complexes containing BBI in animal and human tissues is important in studies of allergenicity of soy protein, in nutritional and toxicological studies, and in pharmacological investigations of anti-carcinogenic effects.
Various immunoassays, including enzyme-linked immunosorbent assay (ELISA) methods have been used for quantitation of protease inhibitors like KTI (Brandon et al., 1987, supra). However, BBI has some features which made the practicality of monoclonal antibody-based immunochemical methods uncertain. As discussed above, BBI is a small molecule of low antigenicity in rabbits. It could not therefore be predicted that an immune response could be generated in other species, such as the mouse. Further, it could not be predicted that BBI could be attached to another molecule for use as a labeled ligand with retention of its antigenic structure. Retention of structure under these circumstances is necessary for synthesis of nonradioactive derivatives with long shelf life suitable for commercially useful assay kits. Retention of antigenic structure is essential to achieve the specificity and sensitivity required of the immunoassays. Further, it is known that some proteins undergo structural changes which affect their antigenicity when adsorbed onto plastic assay plates (S. E. Dierks at al., Molecular Immunology 23: 403-411 (1986)), so it could not be predicted that the antigenic sites of BBI and its complexes with proteases would be readily measurable utilizing solid-phase assay formats.
Retention of antigenic structure would not be sufficient to permit a solid-phase assay. Binding by a small molecule like BBI to the solid phase could leave the antigenic site or sites sterically hindered and unable to bind antibody. Therefore, it could not have been predicted that BBI could be coated on a solid phase for a practical ELISA method. In addition, it was unknown whether a monoclonal antibody with the appropriate specificity and affinity for BBI would also release BBI from the antibody complex under mild conditions to permit a practical affinity chromatography procedure. If BBI could not be released under mild conditions, it would not retain its native configuration, and the method might be useless for purification of BBI or for specific targeting of BBI to a particular site. Further, it was not expect that BBI could be bound by specific monoclonal antibodies and fully retain its ability to interact with a protease, for example, chymotrypsin. The antibody could cause a change in the shape of the chymotrypsin-reactive site or could sterically hinder this site. The ability of an antibody to bind without altering or hindering the chymotrypsin-reactive site would permit targeting BBI to a particular tumor cell, for example, where it could react with a chymotrypsin-like site important in carcinogenesis.